Can an early 24-hour EEG predict the response to the ketogenic diet? A prospective study in 34 children and adults with refractory epilepsy treated with the ketogenic diet

Purpose We examined whether early EEG changes in a 24-h EEG at 6 weeks of treatment were related to the later clinical response to the ketogenic diet (KD) in a 6-month period of treatment. Methods We examined 34 patients with heterogeneous epilepsy syndromes (21 children, 13 adults) and found 9 clinical responders (=50% seizure reduction); this is a responder rate of 26%. We visually counted the interictal epileptic discharge index (IED index) in % during 2 h of wakefulness and in the first hour of sleep (method 1), and also globally reviewed EEG changes (method 2), while blinded to the effect of the KD. Results At group level we saw a correlation between nocturnal reduction of IED-index at 6 weeks and seizure reduction in the follow-up period. A proportional reduction in IED index of 30% from baseline in the sleep EEG, was associated with being a responder to the diet (Pearson Chi-square p = 0.04). EEG scoring method 2 observed a significantly larger proportion of patients with EEG-improvement in sleep in KD responders than in non-responders (p = 0.03). At individual level, however, EEG changes did not correlate very strongly to the response to the diet, as IED reduction in sleep was also seen in 15% (method 1) to 26% (method 2) of the non-responders. Conclusion Nocturnal reduction of IEDs is related to the response to the KD, however in daily clinical practice, an early EEG to predict seizure reduction should not be advised for individual patients.